Over the last 5 years machine learning methods have been used with increasing frequency to analyze fMRI data in a multivariate manner.
The common goal in the studies resorting to these methods is to show that there is information about a variable of interest (e.g. which stimulus is being shown to a subject or which decision the subject will make) in some subset of the data (e.g. voxels in a given location or with particular response characteristics). In this talk I'll introduce the main ideas underpinning a basic analysis of fMRI data with machine learning methods, with an emphasis on showing how the methods match particular scientific questions of interest.